HPLC A Praactical User''''S Guide Part 10 pot

resolution sum is calculated. This process continues with the lowest value of
each new triad being discarded, reflection around the axis joining the best two
points, and a new injection made at the new set of conditions. This technique
will hunt and search until a point (10) is found that meets the search criteria.
The search can be stopped at this point, but there is a danger that only a local
“best” value has been found. If the overall best condition is desired, this final
point can also be discarded, a new point selected at random, and the random
walk can be continued. If the computer continues to return to the previous
“best” point, then it probably represents the true best value within the limits.
Obviously, a limiting maximum number of injections should be set to keep the
computer from wandering around forever.
Although flow rate and solvent composition are the most commonly opti-
mized variables, there is no reason why temperature and other mobile phase
modifiers could not be used. Variables such as mobile phase pH, buffer con-
centration, ion pairing reagents, a chelator’s concentration, or organic modi-
fiers could all be optimized using resolution sums. If the computer can control
the variable UV detector’s wavelength, wavelength and detector sensitivity
settings could both be included as independent variables to be searched and
optimized.
To create a method, you would need a computer-controlled gradient HPLC
system, with an autoinjector or autosampler capable of making repeated
AUTOMATED METHODS DEVELOPMENT 175
Figure 14.3 Random walk optimization.
injections from a large supply of the target solution, an HPLC column, and a
detector. Data acquisition and processing can be done in an integrator and
sent to the computer or can be handled using an A/D card and software
running in the control computer. The system would be set up with sufficient
mobile phase for an overnight run, limits set, and the system allowed to run
unattended overnight. When you come in the next day, the system will be
either still be running chromatograms or the report will be ready with the best
chromatographic conditions on the final printout.

14.5.2 Hinge Point Gradient Development
The development system is usually designed to first try and optimize a fast-
running, two-solvent isocratic separation (variables equal %B and flow rate).
If this cannot be achieved within the run time and expected peak limits, a deci-
sion must be made by the operator as to the next type of development. If your
peaks are nearly separated, you might try making an alpha change by select-
ing a different column and repeating the automated methods development.
Unfortunately, there is no way of making column scouting an automated
procedure.
If your system is a two-pump gradient system, the next step is probably
development of a binary gradient. If you have a multisolvent gradient system,
you usually try to create a binary solvent gradient method before trying to
optimize a three-solvent or even a four-solvent isocratic method in the same
fashion that we optimized a two-solvent isocratic separation.This decision may
just be a case of linear thinking; it is much easier to visualize binary gradient
development than multisolvent isocratic development.
To manually create a binary gradient, a linear gradient is run from 0 to
100%B, the resolution sum calculated, and then a hinge point development is
begun, as discussed in Chapter 12.Automated gradient development works in
a similar fashion; one hinge point at a time is selected and optimized to
improve the separation of compacted peaks by introducing hold before this
area. The hinge point can be entered after operator inspection or at random
time intervals. After resolution is maximized for compacted areas, slope
increases can be introduced at random hinge points to speed run time while
maintaining resolution. Gradient software development is very much a
research science at the moment.
If neither binary gradient nor three-solvent isocratics are successful, some
systems will next try to perform a three-solvent gradient optimization. This
development is very difficult to visualize.Assuming simultaneous optimization
of %B, %C, and flow rate hinge points, it takes a long, computation-intensive

time to carry out. It would be nearly impossible to carry out manually.The key
is continually to use the rule of one: change only one variable at a time and to
carefully select limits for evaluation.
These last changes are probably of academic interest only. Most separations
can be achieved nicely with either a two-solvent isocratic or a binary gradient.
176 AUTOMATION
Most tertiary isocratics in the literature only use a constant level of the third
mobile phase as a polisher. Amines that tend to tail under neutral pH com-
plicate the development. Moving to an end-capped column of adding a fixed
amount of organic modifier will usually fix the problem.Acids can be handled
by going to a lower pH using a fixed amount of acid to buffer pH.
14.6 DATA EXPORTATION TO THE REAL WORLD
Raw data and reports can be stored in the computer’s archival memory, but
they must be transmitted to the real world to be of use. In the simplest case,
they can be displayed on the computer’s monitor in the form of chromato-
graphic curves, tables of data, and reports, or they can be sent to the printer
for printing.They can also be shared with other computers or with other soft-
ware applications for further processing and extraction. To move data out of
the resident software program, they generally have to be translated into some
standard format recognized by other applications.
Laboratory Information Management Systems (LIMS) are computer soft-
ware-based integrators for laboratory reports generation. They gather all the
information on a particular sample, including history, source, supplier address-
ing, data reports from all wet and analytical instruments, and conclusions and
results drawn from this analysis. They receive information from a variety of
inputs, in a variety of formats, and must have inputs for data confirmation and
checking.
14.6.1 Word Processors: .ASC, .DOC, .RTF, .WS, .WP Formats
The simplest of the formats used to transfer data into word processing appli-
cations is the ASCII (.ASC).ASCII is a standard set of 128 binary codes used

by all computers to represent all the characters presented on the normal or
shifted keyboard plus control codes, originally intended for use on teletype-
writers. These code allow us to display lower case leters, capital letters,
numbers, and punctuation marks, but formatting codes for underline,boldface,
and italics are not included in ASCII, and are removed in converting formats.
ASCII files have space-separated code and can be sent out over a modem or
a serial cable to another computer and applications importing ASCII code.
Other word processing formats use higher-level coding in addition to the
ASCII character codes to create proprietary coding specific to that manufac-
turer’s software. Many of these can be recognized and translated by other
writing applications, including the Word Star (.WS) file format and the Word
Perfect (.WP) format. The most commonly used file formats today are the
.DOC format used by Microsoft Word and Rich Text Formatting (.RTF), rec-
ognized by most word processing software and capable of retaining and trans-
mitting formatting information along with the character coding.
DATA EXPORTATION TO THE REAL WORLD 177
14.6.2 Spread Sheets: .DIF, .WK, .XLS Formats
The next type of standard output is the spreadsheet. These file formats use
comma-separated ASCII code, but also add calculation information and
addressing information for the columns and rows they occupy.The simplest of
these are .DIF files, which originated to allow information transfer between
VisiCalc worksheets in the Apple II computer and have been retained as a
standard format. .WK files are Lotus-1,2,3 formats and .XLS are Microsoft
Excel formats that have become spreadsheet standards, allowing transfer of
data, calculations, addresses, and macro programs.
14.6.3 Databases: .DB2 Format
To export data files into a database program, a database file format called
.DB2 was developed in an early PC database, dBase II. Databases are made
up of files, which could be compared to a Rolodex
®

file box full of cards, all
containing the same type of information. The Rolodex
®
card would be equiv-
alent to a database record. Each record has on it a series of entries, fields,in
the same place on each card. To import data into a database record, all the
entries in the report must be matched up with existing fields in the database’s
format. Most software that uses database formats has export/import subpro-
grams that allow you to align fields between the two formats and allow you to
select various ways of determining coding for end-of-file and end-of-record
terminators.
14.6.4 Graphics: .PCX, .TIFF, .JPG Formats
Graphics, the fourth type of export from chromatographic data, is the most
difficult. We can export copies of the monitor screen as bit maps in standard
graphical formats such as .TIFF or .PCX files or in compressed .JPG files, but
much of the fine detail and companion information will be lost.These bit map
files can be manipulated, cleaned up, and labeled in “paint”-type applications,
and then exported into word processing applications. However, the chro-
matogram can no longer be resized and data extraction and integration are no
longer possible. In some graphical applications it is possible to write a printer
format such as .EPS or .HTTP to a file similar to a postscript file, and these
can be used by some applications to resize, rotate, and reprocess the graphi-
cal output.
14.6.5 Chromatographic Files: Metafiles and NetCDF
Chromatographic data file formats are very often in system- and manufac-
turer-specific metafiles. The formats that are used to store these files within an
integrator or data processing unit are usually not designed for export, or they
are designed for export only to other modules by the same manufacturer.They
178 AUTOMATION
may be in a proprietary format, in a compressed storage format, or even gen-

erated under a different computer operating system than in current usage.
Many offer the capability of translating part of their contents to a standard
computer format, but a great deal of information, especially graphical infor-
mation, is lost in the process.
To overcome this problem, a standard chromatographic file format,
NetCDF, was developed and approved by a committee of chromatographic
companies in 1991. It languished for many years until the need to integrate
information from across a laboratory lead to the appearance of LIMS to auto-
mate report generation. This would have been impossible with the babble of
chromatographic information existing only a few years ago.
Every day, data systems are declared obsolete and no longer supported by
their supplier, computer operating systems change and become obsolete, and
hard drive and tape storage systems break down. It quickly becomes obvious
to research laboratories how transient and fragile their archived data files
really are. It is critically important to have access to file translation from these
proprietary formats into a standard format running on modern computer
systems.
DATA EXPORTATION TO THE REAL WORLD 179
15
RECENT ADVANCES IN
LC/MS SEPARATIONS
181
Growth in HPLC systems sales had reached almost replacement level when
adjusted for inflation until about five years ago. The rapid acceleration of the
application of LC/MS systems to solving problems in pharmaceutical research
reversed the trend and then gave it a new upward slope. The pharmaceutical
industry has always been fruitful ground for developing HPLC uses and appli-
cations. LC/MS became the obvious, although expensive, answer for com-
pound identification once atmospheric pressure ionization interfaces matured
enough to provide a robust and reliable bridge between the workhorse HPLC

and the definitive mass spectrometry detector. An additional spurt in systems
sales occurred as proteomics discovered the advantage of using computer-
assisted LC/MS/MS polypeptide fragmentation identification for protein
characterization.
The mass spectrometer detectors place new demands on the HPLC system.
The MS interface requires use of volatile buffers and reagents. Nanospray
interfaces especially benefit from low-volume, high-resolution separations.The
mass spectrometer is a fast response system and benefits from separation
speeds higher than normally supplied by HPLC systems. All of these require-
ments have provided constraints on new development directions for HPLC
systems.
15.1 A LC/MS PRIMER
One of the most important additions to the HPLC arsenal was the
development of the evaporative ionization interface that allowed a mass
HPLC: A Practical User’s Guide, Second Edition, by Marvin C. McMaster
Copyright © 2007 by John Wiley & Sons, Inc.
spectrometer to be use as a detector.The basic LC/MS system (Fig. 15.1) con-
sists of an HPLC pump or gradient components, an injector, and a column
mated to a mass spectrometer through an evaporative/ionizing interface. The
simplest chromatogram produced by this system is similar to a UV chro-
matogram, although possibly with peaks annotated with molecular weights.
The mass spectrometer has the advantage of not only being a universal mass
detector, but also of providing a definitive identification of the compounds
being analyzed. This advantage does not come without difficulties; mass spec-
tral detectors are very expensive compared with other detectors, large com-
puter data storage is required for the mass of information produced, and
compound identification other than molecular weight requires more complex
equipment and considerable interpretation skills. Although prices are coming
down, mass spectral detectors are still primarily research systems costing in
excess of $100,000, with interfaces costing $3–5,000. The high-vacuum pumps

required to run the system have become much more reliable, more compact,
and less expensive, but still require considerable maintenance. Fragmentation
data needed to provide data for structure interpretation as provided by a
GC/MS still requires use of LC/MS/MS systems costing $200,000.
But there are signs that simpler, less expensive LC/MS systems designed
and priced for the general laboratory bench chemist, production facilities, and
quality control laboratories may soon be possible. It remains to seen whether
manufacturers will decide to produce these systems. Older MS systems have
been purchased, attached to HPLC systems equipped with relatively inex-
pensive interfaces, and pressed into service for molecular weight determina-
tion as a $30,000 detector,indicating that the desire and need exists for general
laboratory LC/MS systems. As prices continue to drop and technology
advances work their way out of the research laboratories, the LC/MS will
become a major tool for the forensic chemist whose separations must stand
up in court, for the clinical chemist whose separations impact life and death,
and for the food and environmental chemist whose efforts affect the food we
eat, the water we drink, and the air we breathe.
With this in mind, let us take a look at the design of the LC/MS, its opera-
tion, and the way mass spectral data are manipulated to produce chromato-
graphic information and compound identification. This will be simply an
182 RECENT ADVANCES IN LC/MS SEPARATIONS
Figure 15.1 LC/MS system model.
overview; detailed information is available in LC/MS: A Practical User’s
Guide, listed in Appendix G. Mass spectrometry is a science in itself, but it is
important for the chromatographer to have a working knowledge of its
techniques.
15.1.1 Quadrupole MS and Mass Selection
The mass spectrometer has been around for a long time, with its major shift
into the research laboratory occurring as an outgrowth of the Manhattan
Project during World War II. In the 1960s, a useable GC/MS interface was

developed, but the first commercial HPLC/MS interface did not appear until
the 1970s.A useable atmospheric ionization interface was not developed until
the 1990s because of the problem of seeing compounds in the presence of all
that solvent.
Mass spectrometers work on the principle that a charged ion being pro-
pelled through a curved magnetic field will be deflected inversely proportional
to its molecular mass and proportionally to its charge, allowing us to define an
ion mass term corrected for its charge, m/z. The lighter the mass, the more
deflection that will occur at a given charge. The higher the charge, the more
deflection that will occur at a given mass.
The first research instruments were based on the ungainly magnetic sector
mass spectrometers that used very large permanent magnets to establish the
electromagnetic field and had very slow response times. The accelerated ions
of different masses were detected at different impact points on the detector
plate and mass ratios were measured (Fig. 15.2).
The first useful research instruments were based around the quadrupole
mass spectrometer. Quadrupole mass spectrometers also employ an ion
source, a lens to move the charged ions into the quadrupole mass analyzer
rods, and a detector, all under high vacuum (>10
−5
mmHg). Mass separation is
accomplished in a direct current (dc) quadrupole electromagnetic field applied
acrossed the mass analyzer rods and is modified by a radio frequency (RF)
signal for mass separation and to select and focus the desired mass at the
detector (Fig. 15.3). By sweeping the dc/RF field through a range of frequen-
cies, the quadrupole can be made to focus a series of ions of increasing mass
on the detector, allowing a continuous measurement of m/z through a selected
AMU (atomic mass unit) range (SCAN mode). Alternatively, the quadrupole
can be stepped to specific AMU values in a single ion-monitoring (SIM) mode.
Scan mode is generally more useful when doing qualitative detection, mass

scouting, and in fragmentation studies of unknowns. SIM mode is used for
high-sensitivity detection and quantitation.
Another commonly used type of mass spectrometer is the tandem mass unit,
also referred to as an MS/MS (Fig. 15.4) or a triple quad mass spectrometer.
Originally, this was made up of two or three mass spectrometers used in series.
One MS is used to separated ions,the middle unit is used as a collision chamber
in which selected ions are allowed to impact heavy gas molecules and fragment,
and the last MS is used to separate and measure the fragment ions. In one
A LC/MS PRIMER 183
184 RECENT ADVANCES IN LC/MS SEPARATIONS
Figure 15.2 Magnetic sector mass spectrometer.
Figure 15.3 Quadrupole mass spectrometer.
common MS/MS experiment, the first MS unit is used to separate out a specific
molecular ion and the second MS is used to examine fragmentation daughter
ions that can be used to determine the molecular structure of the original mass
ion by comparison to know fragmentation patterns. Alternatively, the third
quad can be used to scan the fragmentation ions looking for a specific mass ion
to aid in confirming the molecular ion’s identity.
15.1.2 Other Types of MS Analyzers for LC/MS
The quadrupole MS detector was the first, and is still the most common, detec-
tor used for LC/MS, but a number of other mass spectrometers have been
adapted to this application. Both three-dimensional spherical (ITD) and linear
(LIT) ion trap detectors offer tremendous potential for general, inexpensive
LC/MS systems. They both offer the ability to be used as either a mass spec-
tral detector or as a MS/MS detector. The 3D ITD (Fig. 15.5) allows ions to
be trapped in the ion trap where they can be fragmented by heavy gas colli-
sion and the fragments released by scanning the dc/RF frequency of the trap.
The linear ion trap (Fig. 15.6) is essentially a quadrupole detector with an
electrically controlled ion lens at either end. It can trap a much larger volume
of ions in its trap, allowing much higher sensitivity in fragment ion detection

for trace analysis as well as MS
n
-type of experiments in which fragmentation
ions can be trapped and further fragmented to aid in structure studies.
Time-of-flight (TOF) MS detectors (Fig. 15.7) are commonly used in pro-
teomics studies of proteins and protein fragments because this type of detec-
tor can handle and analyze very large molecular and fragmentation ions.
Fourier transform mass spectrometers (FTMS) are being incorporated into
commercial LC/MS systems and offer the advantage of being nondestructive
detectors that can trap and repeatedly analyze the same sample in order
A LC/MS PRIMER 185
Figure 15.4 Quadruple LC/MS/MS system.
186 RECENT ADVANCES IN LC/MS SEPARATIONS
Figure 15.5 Ion trap detector.
Figure 15.6 Linear ion trap MS analyzers. (Courtesy of Applied Biosystems.)
Figure 15.7 Time-of-flight MS analyzer.
to greatly increase analysis sensitivity for things like accurate mass
determinations.
15.1.3 LC/MS Interfaces
The first modern LC/MS interface was a thermospray interface introduced in
1983 that allowed introduction of column effluent at 1.0–1.5mL/min. The
mobile phase had to be highly aqueous and contained large amounts of
volatile buffer to induce chemical ionization. The mobile phase was forced
through an electrically heated capillary and out through a fine orifice into the
high vacuum source of the mass spectrometer. The entrance capillary often
had to be heated to >200°C and the requirement for >100mM volatile buffer
often lead to sample decomposition and orifice plugging.Roughing pumps had
to be added to the system to remove the large amounts of solvent and volatile
buffer release in the ionization process.
Today, the two most common LC/MS interfaces are atmospheric pressure

ionization interfaces, electrospray (ESI) and ion spray (ISI).Electrospray (Fig.
15.8) and its subtype, nanospray, are recommended for use with proteins and
highly polar or ionized compounds. They are very soft ionization, concentra-
tion-dependent techniques that result in very little fragmentation and often
produce multiply charged molecular ions.
Ionization is accomplished in the electrospray interface by passing the
HPLC effluent down a heated metal capillary tube along which an electric
charge differential is applied. The evaporating liquid sprays out of the tube
end as charge droplets rapidly decreasing in size. A gas nebulizer often
A LC/MS PRIMER 187
Figure 15.8 Electrospray interface.
encloses the capillary and aids in droplet evaporation. Electrospray is best
done at reduced microliter/minute flow rates.
Proteins can acquire multiple positive charges at basic amino acids such as
lysine to form multicharged molecular ions. Since the MS analyzer separates
ions on the basis of m/z, or mass divided by charge, mass spectrometers with
an operating range of 0–2,000amu can still detect proteins with masses up to
100,000amu if they have 10–50 charges per molecule. Deconvolution software
can resolve this peak charge envelope developed by a single protein to allow
calculation of the protein’s molecular weight.
An ISI is used with less polar effluents and is the workhorse for standard
HPLC systems since it can take flow rates up to 2mL/min.It is most commonly
used to produce intact molecular ions for molecular weight determinations,
but it can be set up with an ion repeller to cause fragmentation that can
provide preliminary compound identification and structural information.
The ISI interface (Fig. 15.9) also uses a nebulizing inert gas to entrain and
break up the eluant stream into small droplets that are sprayed across a
coronal discharge needle operated at about 25KV to ionize the shrinking
droplets.The impactor plate is equipped with a charge opposite to that applied
to the coronal needle to draw the charged ion to the mass spectrometer

entrance. Again, the nebulizer capillary may be heated to aid evaporation or
an oppositely charged plate may draw the charged droplets into a heated tube
before they enter the mass spectrometer inlet. The ISI is a mild ionization
source and generally produces only a single, molecular ion per compound
unless the molecular ion is very unstable. Fragmentation does not usually
occur as it does in an ESI source of a GC/MS system. A MS/MS-capable
system must be used to produce daughter ion fragmentation for structural
studies.
188 RECENT ADVANCES IN LC/MS SEPARATIONS
Figure 15.9 The ion spray interface.
Most modern LC/MS systems are equipped with replaceable ESI and ISI
interfaces, depending on the types of compounds expected in the effluent.
Agilent Technologies has recently released a multimode source capable of ion-
izing all compound types with only minor loss in sensitivity compared with
dedicated sources. If this interface meets its initial claims and can be retrofit-
ted to other LC/MS systems, it could rapidly replace either of the other two
interfaces for general applications instruments.
Another interface commonly used for connecting HPLC to a mass spec-
trometer is not a true in-line interface. It is a robotically controlled spotter
plate system for collecting samples from the HPLC to be injected into the
MALDI time-of-flight laser ionization mass spectrometer for analyzing pro-
teins and large peptides.The effluent sample dropped in the plate well is mixed
with an ionization matrix already present, solvent and volatile reagents are
evaporated, and the plate is then placed into the injector target and blasted
with a pulsed laser to volatilize and ionize sample into the atmosphere of the
interface where it can be drawn into the mass spectrometer.
15.1.4 LC/MS Computer Control and Data Processing
The mass spectrometer like the diode-array UV detectors produces a three-
dimensional array of time-voltage-spectral information and requires consid-
erable computer power to handle both MS scanning control and data

processing.A simple general-purpose ISI-LC/MS is designed to separate com-
pounds and present a chromatogram with peaks identified by retention times
and molecular weights. The system computer must start/stop the system, scan
the MS mass range, identify the molecular weight of the major molecular ion,
and display the total ion chromatogram (TIC) annotated with the major peak
molecular weight.
The MS detector is more sensitive and requires much less sample than UV
detection. The LC/MS system is often designed either with an in-line sec-
ondary detector or with a splitter system before the interface and a second
detector so that additional information can be obtained on the sample. Figure
15.10 compares the chromatographic signals from a variable UV detector and
the full-scan total ion chromatogram (TIC) from a MS detector run in series
for detection of an adhesive extract.
A more complicated task faces an ESI-LC/MS designed to separate and
determine the molecular weights of proteins. The proteins have to be sepa-
rated by the column, the multiply charged ion envelope must be measured,
and deconvolution calculations made to determine the molecular weights of
the separated protein(s). Often, there are partially resolved proteins mixtures
and the overlapping peak envelopes must be resolved to determine molecu-
lar weight for both components present.
Working with a LC/MS/MS system adds further degrees of complication.
The system computer must control both MS units either in a scanning or in a
A LC/MS PRIMER 189
single ion mode. The data array must be examined for fragmentation patterns
for daughter ion identification, hopefully using a spectral library database.
Anything but the simplest routine procedure becomes a mass spectrometry
research project.
190 RECENT ADVANCES IN LC/MS SEPARATIONS
Figure 15.10 Comparison of (a) UV and (b) TIC full-scan MS chromatograms of adhesive
extract. (From Tiller et al., Copyright © 1997 John Wiley & Sons, Limited. Reproduced with

permission.)
15.2 MICROFLOW CHROMATOGRAPHY
Two factors are driving the market for precise, very-low-flow HPLC pumping
systems: extremely limited sample sizes in biotechnology and the electrospray
and nanospray interfaces that are concentration and flow-rate dependent. It
is very difficult to get precise flow and gradient formation from pumps that
have a 5- to 10-mL plunger displacement, even using 3200-step stepper motor
drives.This has forced manufacturers to resurrect a very old concept from the
earliest days of HPLC, the syringe pump.
A syringe pump is a cylinder equipped with a tight-fitting piston and inlet
and outlet check valves. Opening the inlet check valve and pulling back on the
plunger fills the cylinder. Closing the inlet valve, opening the outlet valve and
driving the plunger forward initiates flow. Syringe pumps generate smooth,
pulse-free flow, but their delivery solvent volume is limited to the capacity of
their heavy-walled cylinder. In practice, this translates to an upper volume
limit of about 250mL, which is insufficient for most standard HPLC separa-
tions. Beyond this volume, pressure-resistant cylinder wall armoring became
cost prohibitive.
Micro- and nano-HPLC systems (Fig. 15.11) rely on small-diameter and
capillary columns packed with high-efficiency packing materials and very slow
flow rates to produce concentrated solutions and sharp chromatography peaks
to feed electrospray interfaces for mass spectrometers.
Solvent volumes per run do not exceed the capacity of the syringe pump(s)
because of the extremely low flow rates. Injection size is limited in these
systems by incorporating a fixed-volume injector loop within the body of the
MICROFLOW CHROMATOGRAPHY 191
Figure 15.11 Syringe pump micro-HPLC.
injector. Decreased diameter and volume of tubing and connectors further
increases the efficiency of these systems.
15.3 ULTRAFAST HPLC SYSTEMS

A different philosophy is taken with the ultrafast chromatography systems,
such as the ACQUITY
R
ultraperformance liquid chromatography (UPLC)
system that dominated the pharmaceutical LC/MS market after its introduc-
tion in 2005. These systems are also designed to deliver highly concentrated
effluent quickly to a mass spectrometer. But, the ultrafast systems use very
small particle size HPLC-packed columns run at very high system pressures.
Silica has always had the mechanical strength to handle pressures to about
15,000psi.The problem has been in sealing the system to resist leakage.
When first introduced, the only column available for the ACQUITY
R
system was a 1.7mL C
18
hybrid silica, but an expanded line of C
18
,C
8
, and
phenyl bonded-phase columns are now offered for this system. As column
diameters moved from 10 to 5 to 3m, it was found that the resolution loss at
high mobile phase flow rate began to level off. Particles with 1–3m diameters
show very little if any resolution loss at very high flow rates. Of course, very
small diameter particles still show very high back-pressures at high flow rates,
require very small pore frits to keep them in the column, and are very easily
contaminated with particulate matter in the mobile phase. UPLC and nano-
UPLC systems are now being sold that are optimized for LC/MS operation at
12,000psi back-pressure and flow rates up to 10mL/min flow rate with a cor-
responding decrease in separation times.
15.4 CHIP HPLC SYSTEMS

Another interesting HPLC technology is the miniaturized HPLC-on-a-chip
system. An outgrowth of microfluidics technology used in inkjet printing and
automated bioanalyzers, it consists of tiny closed channels etched onto a glass
or plastic microchip. This system is designed to marry into a microfluidics
nanospray interface on a second chip that will spray directly into a mass spec-
trometer. Used in an interface mounted on a mass spectrometer, they provide
an easily replaced works-in-a-drawer solution to column and nozzle plugging
while minimizing extracolumn problems normally associated with tubing and
fittings.
Agilent Technologies currently offers commercial nanospray HPLC-Chip
R
and HPLC-Chip Cube
R
MS and MS/MS systems. Chips come with standard
C
18
packing or can be custom packed with standard HPLC column
materials. They are aimed at proteomic labeling studies, and small molecule
separations.
192 RECENT ADVANCES IN LC/MS SEPARATIONS
15.5 STANDARDIZED LC/MS IN DRUG DESIGN
Standard or generic LC/MS methodology was developed by pharmaceutical
companies to provide rapid screening and a common informational database
throughout a company’s activities. Used from drug development screening
through process monitoring and metabolite and degradation studies on com-
pounds from almost any media, such as urine, plasma, or reaction mixtures,
only the sample preparation will vary. No attempt is made to optimize the
chromatographic separation for individual compounds or classes of com-
pounds. A standard 5–20-min linear reverse-phase gradient from 5% organic
in water to 95% organic at neutral pH is run on a specific C

18
column type
from a single manufacturer. The mass spectrometer is run in scan mode over
a standard m/z range with a time delay to ignore the solvent spike and the
retention time and molecular ion mass is determined for each compound.
The advantage of generic LC/MS run conditions is that it allows the prepa-
ration of an LC/MS separation database that can be referenced for compound
mixtures from anywhere in the development and manufacturing process cycle.
It trades off resolution for consistency,speed,and a decrease in methods devel-
opment times. It permits creation of a computer-searchable database of infor-
mation for all of the compounds being investigated in the company. The mass
spectrometer provides sensitivity and resolution gain as well as information
on retention times and molecular weights.
When I first saw the linking of combinatorial chemistry with generic LC/MS
methods I was appalled to the depth of my traditional chemist’s soul. It looked
like they were using the power of the mass spectrometer resolution to try to
fix bad chemistry and bad chromatography. But when I studied the technique
I realized it used the same resolution, speed, and load triangle that we take
advantage of in doing preparative chromatography. There we sacrifice resolu-
tion to gain load and speed. In generic chromatography, they trade off reso-
lution for speed and analytical compatibility across the drug discovery and
development process.
STANDARDIZED LC/MS IN DRUG DESIGN 193
16
NEW DIRECTIONS IN HPLC
195
Separation speed and ease of use seem to be the primary factors driving
changes in HPLC instrumentation. Resolution efficiency and stationary phase
stability, especially at high pH, are the primary factors affecting current
changes in column technology.

16.1 TEMPERATURE-CONTROLLED CHROMATOGRAPHY
Temperature has often been suggested as a useful control variable for HPLC
to make a changes and to speed equilibrations leading to faster separations.
The problem has been that both bonded-phase hydrolytic cleavage and solu-
bility of silica in aqueous solvents are accelerated at elevated temperatures.
Mobile phase boiling within the column can cause bubble formation and vapor
locking if the critical point of the solvent is exceeded. Finally, thermal-labile
compounds can suffer degradation at elevated temperatures.
Many of these problems disappear when a hybrid-silica column or a
zirconium-based HPLC column is selected. Hybrid-silica packing material has
an organic skin coating and protecting the silica surface. Advanced zirconium
bonded-phases are chemically bound directly to the zirconium surface, which
does not dissolve in aqueous solvents. ZirChrom has released a newsletter
(vol. #5) 2001 (see Appendix G) claiming a 12-fold decrease in separation time
without loss in resolution using an efficient column heater.The technique ben-
efits from using narrower columns to avoid temperature gradients in the
column and mobile phase preheating before entering the column. Selecting a
HPLC: A Practical User’s Guide, Second Edition, by Marvin C. McMaster
Copyright © 2007 by John Wiley & Sons, Inc.
relatively high boiling mobile phase such as aqueous acetonitrile allows
separations at 120°C because the elevated pressure in the column raises the
boiling point. The hybrid-silica columns show some of the pH stability of the
chemically bound zirconium columns but probably would show decreased
operating lifetimes due to increased packing material solubility.
16.2 ULTRAFAST CHROMATOGRAPHY
With large diameter HPLC packings, efficiency decreases rapidly at higher
flow rate, as does back-pressure. Five- and 10-mm spherical packings show a
maximum efficiency at 1–1.5mL/min flow rate. However, when 3-mm packing
appeared they showed little loss of efficiency at 3.0mL/min; the back-pressure
increases dictated the use of shorter columns, trading efficiency for shorter run

times. Use of finer particles required the use of finer bed support filters and
increased the danger of plugging by particulates in the injected sample.
The UPLC Acquity
R
system mentioned in Chapter 15 takes this flow
rate–resistant efficiency gain another step using a 1.7-mm hybrid-silica packing
material run at back-pressures up to 12,000psi to provide flow rate decreases
of 3–5-fold. Silica has always had the capability to resist high back-pressures;
many traditional HPLC columns are packed with 20,000psi pumping systems.
But, system components from pumps, injectors, and fittings all had to be
redesigned to resist the extra pressure. Extra column volumes had to be
reduced to avoid losing the increase column efficiency.A column heater is built
into the system to take advantage of the increased hydrolytic stability of the
hybrid-silica column in speeding flow rates. Obviously, a limit exists as to how
much farther you can reduce the particle size of the packing material and
still keep it with in a column and pump solvent through it at an acceptable
back-pressure.
16.3 MONOLITH CAPILLARY COLUMNS
In the first edition of this book, I forecast that the ultimate HPLC column
would be a “wall-bonded” capillary column that would avoid the voiding and
back-pressure problems seen with packed columns.A new type of column, the
monolith silica column, recently emerging from research laboratories very
closely fits this description. A monolith column has a honeycomb foam of
silica, which is bonded with an organic separating phase, completely filling the
inside of the column.
One way of preparing the silica honeycomb is to mix tetramethoxysilane in
a porogenic solution of polyethylene glycol and urea in the presence of acetic
acid. This mixture is poured into a capillary tube treated with sodium hydrox-
ide. The tube is first heated to hydrolyze the urea to form pores and then
further heated to burn off organics leaving a continuous silica foam skeleton

196 NEW DIRECTIONS IN HPLC